Navigating the American healthcare system — informed consent

Doctor: I recommend taking [specific medication] to treat some of the concerns you came into for today. What are the main questions you have about this drug?

Patient: Yes doctor, I do not want any side effects. What are the side effects?

D: Great question, all interventions have trade offs, and some of the potential trade offs of medications are possible side effects. The most common side effect is headache, which occurs at rates similar to placebo. Most people do not notice much improvement at first, which can be disappointing. Other common side effects include dry mouth, GI upset, and nausea. I usually ask patients to bear with this side effects at first, if they occur, because they often improve within a few days. Rarely, more rare than the probability of being struck by lightning on a clear day, people can be afflicted with [very unwanted side effect]…

P: Oh doctor, I do not want [unwanted side effect], is there anything else you can prescribe?

D: Yes, but please be aware that every possible drug I might prescribe for this problem has a similar risk for [unwanted side effect]. However, this is not a reason to avoid prescribing these drugs, but rather information so that you know what to do if you see [unwanted side effect].

P: Well now that you mention it, I really do not want headache either. I had some bad headaches in the past and I would rather just keep [suicidal depression and crippling anxiety] than get headaches and [unwanted side effect].

Hopefully, for the patient, the conversation does not end there.

Probability

Life has great uncertainty, which frequently leads to negative emotional states. Certainty provides temporary relief from negative emotional states such as worry and sadness. If we lose something, we may comfort ourselves by focusing on what certainly remains. Forced change is terrifying. Fear is incredibly useful if it helps us to focus on solving the scary problem. Sometimes avoiding the scary thing is best, but just as often confrontation is optimal.

Probability is the knowledge shortcut for identifying if we should avoid or tangle with the scary thing. Sometimes society makes this decision for us. In the USA, about 1 in 100 people you know will die from injury related to motor vehicle. It is a big country, and motors help the economy move. We as a society of ~300 million seem to accept a 1% lifetime risk of death from driving. If you - yes you - wear a seatbelt then your risk of dying in a collision is cut in half. (Little known fact, drunk walking is very dangerous too).

Back to biology for a moment. All medicinal chemicals, all recreational chemicals, all foods, all ingestibles have a technical risk of DEADLY ALLERGIC REACTION. Albeit the risk of death is infinitesimally small for most common foods/drugs.

The lifetime odds of being struck by lightning are around 1 in 15,000. That means for every 15,000 people you know, on average one will be struck by lightning in their lifetime. Most of us keep a social circle of a couple hundred people at most, so most people will know a motor vehicle victim but not a lightning victim.

Does discussing something rare but possible make it more likely to happen? Many people believe so. In psychiatry we call it magical thinking. In layman’s terms it’s called superstition.

Informed

Discussion of benefits/risk requires discussion of probability. The phrase “the benefits outweigh the risks,” shall be amended as thus: “the probable benefits outweigh the possible harms.”

What are the probable benefits of taking lisinopril? Decreased blood pressure and decreased heart disease. What are the possible harms? Dry cough, allergic reaction, harmful to pregnancy, angioedema — this resembles allergic reaction but it is a different mechanism. Lisinopril is so effective at reducing blood pressure we measure how much it reduces blood pressure, not if.

Medical statisticians have distilled probable benefits vs possible harms into a number system. This system is known as the number needed to treat (NNT) and the number needed to harm (NNH).

Quick aside, the famous phrase, “first, do no harm,” is not a line of the Hippocratic oath.

NNT refers to a concept: if a group of people have a disease, and one-by-one each person is given an intervention, on average how many people need receive the intervention before ONE is treated successfully?

I will put it another way: a group of infinite people with a tender broken arm line up neatly in a queue. A nurse walks down the line and gives an injection of a pain relief medicine to each person, but pauses and asks, “feel better?” after each administration. If the tender broken arm feels no better she moves on to the next. As soon as someone says, “ah, that’s better!” she stops and counts how many people she got through before ONE felt better. (This example assumes an even distribution of people who will benefit, but the line of people is easier to visualize than averages and percentages).

If the NNT = 1 then each patient who received the intervention improved. If the NNT = 10 then only about 10% of patients who receive the intervention will improve. If NNT = 100 then only about 1% of patients who receive the intervention will improve. A respectable NNT for a drug is 5 or less.

The NNH is calculated the same way, but it measures an identifiable harm. For example, if one out of every 50 people who receive the medicine go on to develop kidney failure, then the NNH = 50. Harms have a lot of variety, so mild, moderate, and severe harms may be grouped together.

The NNT had better be much lower than the NNH in almost all scenarios. If that does not make intuitive sense then double check that you have understood the concept.

Consent

Researchers collect the data, statisticians analyze the data, doctors learn the statistics, and patients trust the doctor to inform them. If there is any error in the chain, then the patient is punished. The patient cannot become the researcher, statistician, and doctor all at once. There will always be uncertainty.

I am aware of the drugs that have a high NNT and relatively low NNH. Some of them are drugs that feel good to take — these are drugs not worth taking but difficult to convince the patient once he or she has started. I am aware of the invasive interventions with high NNT and high NNH — these interventions are lucrative for business but neutral to the patient. I am aware of the healthcare charlatans with treatments that have infinitely high NNT and finite NNH — this is risk without any probable benefit beyond placebo.

Consent to be treated, given or denied, is implicated in the trust the sufferer has for the interventionist. This talk of probability and statistics might as well be mumbo jumbo. If the trusted surgeon says she needs to operate then the trusting patient obliges in almost all circumstances. In psychology, blind trust in an individual who is perceived as trustworthy is known as the halo effect. Fortunately, trusting a surgeon is correct so frequently that it makes national news headlines when it was incorrect to trust.

Summary: Consent to receive an intervention is most often granted based on trust; trust is based on emotions rather than probability; a sufferer is captive and beholden to those offering relief.

To whom should we grant consent to be treated? To those who understand and discuss nuance, risk, and probability. In other words, those who are informed and can communicate the information.

Certain institutions do most of the vetting to identify who is informed: Medical boards, specialty boards, the DEA, hospitals. The institutions are imperfect but vastly superior to the alternatives.

-Doc